@inproceedings{govindan-etal-2026-identifying,
title = "Identifying Temporal Features within Transcoders for Time Sensitive Factual Recall",
author = "Govindan, Sanjay and
Pagnucco, Maurice and
Song, Yang",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1442/",
doi = "10.18653/2026.findings-emnlp.1442",
pages = "30761--30782",
ISBN = "979-8-89176-462-0",
abstract = "Large Language Models (LLMs) suffer from temporal misalignment, often due to the contradictory nature of their training corpora. While current mitigation strategies rely on computationally expensive fine-tuning or context-heavy retrieval augmented generation (RAG), the internal mechanisms governing time sensitive recall remain under-explored. Unlike prior studies that identify temporal components such as attention heads and MLP layers, we provide the first feature level map of temporal recall by isolating individual MLP features via transcoder circuit tracing. We identify three node categories (common temporal, common to the year and chrono-semantic) which interact to generate a temporal filter during factual recall. By analysing Gemma 2 2B, LLaMA 3.2 1B, and Qwen3-4b, we show that these features do not follow a simple linear pipeline but represent time through a parallel and mixed syntactic-semantic interplay across layers. We additionally discover a class of higher-layer temporal components invisible to existing EAP-IG methods, establishing transcoders as a more complete lens for temporal interpretability for time sensitive factual recall. These findings present MLP components for potential targeted interventions in time-sensitive factual recall. Our code can be found at https://github.com/sg-sy/timewarp-temporaltranscoder."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="govindan-etal-2026-identifying">
<titleInfo>
<title>Identifying Temporal Features within Transcoders for Time Sensitive Factual Recall</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sanjay</namePart>
<namePart type="family">Govindan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maurice</namePart>
<namePart type="family">Pagnucco</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Large Language Models (LLMs) suffer from temporal misalignment, often due to the contradictory nature of their training corpora. While current mitigation strategies rely on computationally expensive fine-tuning or context-heavy retrieval augmented generation (RAG), the internal mechanisms governing time sensitive recall remain under-explored. Unlike prior studies that identify temporal components such as attention heads and MLP layers, we provide the first feature level map of temporal recall by isolating individual MLP features via transcoder circuit tracing. We identify three node categories (common temporal, common to the year and chrono-semantic) which interact to generate a temporal filter during factual recall. By analysing Gemma 2 2B, LLaMA 3.2 1B, and Qwen3-4b, we show that these features do not follow a simple linear pipeline but represent time through a parallel and mixed syntactic-semantic interplay across layers. We additionally discover a class of higher-layer temporal components invisible to existing EAP-IG methods, establishing transcoders as a more complete lens for temporal interpretability for time sensitive factual recall. These findings present MLP components for potential targeted interventions in time-sensitive factual recall. Our code can be found at https://github.com/sg-sy/timewarp-temporaltranscoder.</abstract>
<identifier type="citekey">govindan-etal-2026-identifying</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1442</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1442/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>30761</start>
<end>30782</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Identifying Temporal Features within Transcoders for Time Sensitive Factual Recall
%A Govindan, Sanjay
%A Pagnucco, Maurice
%A Song, Yang
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F govindan-etal-2026-identifying
%X Large Language Models (LLMs) suffer from temporal misalignment, often due to the contradictory nature of their training corpora. While current mitigation strategies rely on computationally expensive fine-tuning or context-heavy retrieval augmented generation (RAG), the internal mechanisms governing time sensitive recall remain under-explored. Unlike prior studies that identify temporal components such as attention heads and MLP layers, we provide the first feature level map of temporal recall by isolating individual MLP features via transcoder circuit tracing. We identify three node categories (common temporal, common to the year and chrono-semantic) which interact to generate a temporal filter during factual recall. By analysing Gemma 2 2B, LLaMA 3.2 1B, and Qwen3-4b, we show that these features do not follow a simple linear pipeline but represent time through a parallel and mixed syntactic-semantic interplay across layers. We additionally discover a class of higher-layer temporal components invisible to existing EAP-IG methods, establishing transcoders as a more complete lens for temporal interpretability for time sensitive factual recall. These findings present MLP components for potential targeted interventions in time-sensitive factual recall. Our code can be found at https://github.com/sg-sy/timewarp-temporaltranscoder.
%R 10.18653/2026.findings-emnlp.1442
%U https://aclanthology.org/2026.findings-emnlp.1442/
%U https://doi.org/10.18653/2026.findings-emnlp.1442
%P 30761-30782
Markdown (Informal)
[Identifying Temporal Features within Transcoders for Time Sensitive Factual Recall](https://aclanthology.org/2026.findings-emnlp.1442/) (Govindan et al., Findings 2026)
ACL